Experimental Validation of Decentralized Affine Transformation

Multi-Agent Drone Swarm Coordination Using Leader-Follower Framework

Garegin Mazmanyan, Hossein Rastgoftar
September 2025
ROS2 Python Decentralized AT Crazyflie 2.1 Motion Capture System Multi-Agent Systems

Experimental Demonstration

Experimental validation of decentralized affine transformation with Crazyflie 2.1 quadcopters. The video demonstrates the drone swarm executing coordinated maneuvers while maintaining formation and avoiding obstacles in a constrained environment.

Project Overview

This research presents an experimental validation of decentralized affine transformation (AT) in multi-agent systems using teams of mini-quadcopters. The AT framework enables an agent team to safely navigate constrained, obstacle-rich environments while allowing aggressive changes in inter-agent distances, which are formally characterized through the decomposition of the AT transformation matrix.

The project focuses on two-dimensional AT, formulated as a decentralized leader-follower problem. In this formulation, three leader quadcopters are positioned at the vertices of a triangle, while all follower quadcopters remain within the triangle. The leaders know the desired trajectories prescribed by the AT, whereas the followers do not. Instead, the followers infer their trajectories through local communication governed by fixed communication weights determined by the initial spatial configuration of the team.

Methodology

Decentralized Leader-Follower Framework

The decentralized affine transformation approach employs a leader-follower architecture where:

  • Leader Agents: Three quadcopters positioned at triangle vertices that follow prescribed AT trajectories
  • Follower Agents: Quadcopters that remain within the leader triangle and infer their trajectories through local communication
  • Communication Weights: Fixed weights determined by initial spatial configuration, enabling decentralized coordination
  • Asymptotic Convergence: Mathematical guarantees for trajectory convergence without centralized control

Experimental Setup

The experiments utilized the Decentralized AT framework with Crazyflie 2.1 mini-quadcopters and a motion capture system for precise position tracking. The setup enabled real-time validation of the decentralized AT algorithm in obstacle-laden environments, demonstrating the framework's capability to safely guide multi-agent teams through complex scenarios.

Experimental Results

3D trajectory visualization of drone swarm executing affine transformation

Figure 1: 3D trajectory visualization showing the drone swarm executing decentralized affine transformation. The leader quadcopters (positioned at triangle vertices) follow prescribed trajectories while follower agents maintain formation through local communication.

Tracking error analysis demonstrating asymptotic convergence

Figure 2: Tracking error analysis demonstrating the asymptotic convergence of the decentralized AT algorithm. The results validate the theoretical predictions and show successful trajectory following with minimal error.

Key Findings

  • Asymptotic Convergence: Experimental results validate the asymptotic convergence of decentralized AT, confirming theoretical predictions
  • Safe Navigation: Successfully demonstrated capability to guide multi-agent teams through obstacle-laden environments
  • Aggressive Maneuvers: Enabled significant changes in inter-agent distances while maintaining formation integrity
  • Decentralized Control: Achieved coordination without centralized control, relying solely on local communication
  • Scalability: Framework scales effectively with team size while maintaining performance

Technical Contributions

Decentralized Architecture

Novel leader-follower framework enabling autonomous coordination without centralized control

Mathematical Validation

Rigorous experimental validation of asymptotic convergence properties

Safe Navigation

Demonstrated safe operation in constrained, obstacle-rich environments

Formation Flexibility

Support for aggressive changes in inter-agent distances during operation

Publication

Experimental Validation of Decentralized Affine Transformation

Garegin Mazmanyan, Hossein Rastgoftar

arXiv preprint arXiv:2509.10791, September 2025

Subjects: Systems and Control (eess.SY)